Gunes E.D.,Koc University |
Aksin O.Z.,Koc University |
Ormeci E.L.,Koc University |
Journal of Service Research | Year: 2010
Cross-selling attempts, based on estimated purchase probabilities, are not guaranteed to succeed and such failed attempts may annoy customers. There is a general belief that cross-selling may backfire if not implemented cautiously, however, there is not a good understanding of the nature and impact of this negative reaction or appropriate policies to counter-balance it. This article focuses on this issue and develops a modeling framework that makes use of a Markov decision model to account for negative customer reactions to failed sales attempts and the effect of past contacts in managing cross-selling initiatives. Three models are analyzed, where purchase probabilities are affected from customer maturity, the number of failed attempts since the last purchase, or both. The analysis shows that customer reactions to cross-sell attempts make the purchase probabilities endogenous to the firm's cross-selling decisions; hence, the optimal cross-selling policy becomes a function of customer state. The results highlight the role that the cost of excessive cross-selling (direct as well as in the form of customer reactions) plays in optimal policies. Cross-sell data from a retail bank illustrate in what context the modeling framework can be applied and underline the importance of customizing cross-sell policies to individual customers. © The Author(s) 2010.
Koc G.,KocSistem |
Yegin K.,Yeditepe University
International Journal of Distributed Sensor Networks | Year: 2013
Seismic sensors are invaluable for intruder detection and perimeter security. In a typical wireless sensor network application of seismic sensors, the units are battery powered and low power consumption becomes critical while fulfilling system requirements. Although many systems utilize 24-bit ADC for seismic signal processing, we employed 12-bit ADC for low power consumption. Because of this relatively low resolution ADC, preamplifier and filters require careful hardware design. We employ bidirectional T-type filtering, noise reduction, and distributed filtering between gain stages before the signal is input to ADC. The proposed design was verified with measurements. Seismic data signals due to footsteps at varying distances were successfully measured. © 2013 Gokhan Koc and Korkut Yegin.
Agency: Cordis | Branch: FP7 | Program: CP | Phase: FI.ICT-2011.1.8 | Award Amount: 19.54M | Year: 2013
Insights gained in FI-PPP Phase 1 emphasize the need for novel ICT solutions that allow radical improvements for collaboration in business networks. Primary sectors demanding such solutions are Agri-Food and Transport & Logistics industries: several actors (incl. enterprises, authorities, service providers) need to exchange information & communicate across org. borders to conduct business. Current ICT solutions limit this to mostly manual efforts, therewith tremendously hampering business efficiency, effectiveness & sustainability.Leveraging on outcomes of two complementary Phase 1 use case projects (FInest & SmartAgriFood), aim of FIspace is to pioneer towards fundamental changes on how collaborative business networks will work in future. FIspace will develop a multi-domain Business Collaboration Space (short: FIspace) that employs FI technologies for enabling seamless collaboration in open, cross-organizational business networks, establish eight working Experimentation Sites in Europe where Pilot Applications are tested in Early Trials for Agri-Food, Transport & Logistics and prepare for industrial uptake by engaging with players & associations from relevant industry sectors and IT industry.FIspace will extensively utilize GEs and Development Kit from FI-PPP Core Platform, validating its capabilities, openness & versatility for realizing future B2B collaboration solutions. FIspace will be open; other FI-PPP projects and external users & solution providers can easily use, test & exploit it. It supports a future business model where external solution providers, esp. SMEs, can provide additional, novel, and disruptive solutions into the FIspace ecosystem. With this, FIspace will prepare for establishment of a future standard for cross-organizational business collaboration leapfrogging pressing challenges arising in industry, exploit FI technologies developed in FI PPP, and lay foundation for industrial uptake & innovation enablement planned for FI-PPP Phase 3.
Ergen E.,KocSistem |
21st World Congress on Intelligent Transport Systems, ITSWC 2014: Reinventing Transportation in Our Connected World | Year: 2014
Possible accidents caused by the intersections with no traffic control can be prevented by the use of modules integrated to Intelligent Transportation System. Vehicles moving on linear roads can request a data from the map module for the intersections in close range and on their way. Vehicles can share their dynamic data in real-time with the system as well as with the other vehicles in the system. System gather all other cars' information and evaluate those information against its own information and warn the driver about a possible risk of a collision. There are some rules and filters added to the system so system doesn't bother drivers with unnecessary information. Prerequisites for the system to warn the driver are; if vehicles are about to intersect in a predefined time period, and if the vehicles are above the certain speed limit. If vehicles are moving on roads where the intersections are by passed with overpasses and underpasses or there are certain road conditions that force the vehicles to slow down prior to meeting then no warning message is generated. Designed system is proven to be successfully producing warning messages when there is only one intersection between the vehicles and vehicle or vehicles exceed the defined parameters in a hazardous way and there I a risk of a collision.